Making "Weed Maps": Using Mapping to Evaluate (in)Equity in Cannabis Retail Outlet Availability in the United States

Abstract

As of 2023, twenty-four states and Washington D.C. have legalized the production and sale of recreational cannabis for adult use. The establishment of a legal marketplace is followed by the opening of storefront locations from which adults can legally purchase these products. However, there has been minimal inquiry into cannabis retailer location or if patterns of cannabis retail density differ by neighborhood characteristics. The purpose of this dissertation was to examine the distribution of recreational cannabis retailers across early- and later-adopting states, identify the association between neighborhood advantage and disadvantage and cannabis retail density, and describe how cannabis retailer location relates to neighborhood alcohol and tobacco outlets.

Study 1 assessed the association between structural neighborhood advantage or disadvantage with the presence of one or more cannabis retailers in a census tract. Neighborhood advantage and disadvantage was measured using quintiles of the (1) Neighborhood Deprivation Index and the (2) Index of Concentration at the Extremes (ICE). In the adjusted model, tracts in the most deprived quintile had 2.12 times the odds of having at least 1 cannabis retailer compared to the reference group of the least deprived census tracts. Q1 tracts with greatest concentration of Black residents or Hispanic residents, had 1.51 and 1.92 times the odds of having a cannabis retailer compared to Q5 tracts with the greatest concentration of White residents. Q1 tracts with greatest concentration of low-income Black or Hispanic residents had 2.35 and 2.47 times the odds of having a cannabis retailer compared to tracts with the greatest concentration of high-income White residents.

The purpose of Study 2 was to identify the relationship between structural neighborhood advantage and disadvantage with different measures of cannabis retail density (count per tract, cannabis retail density per 1000 population, cannabis retail density per square mile, and cannabis retail density per 10 km of roadway). Quintiles of ICE were used to represent structural neighborhood disadvantage and advantage, comparing iv Income (low-income vs high-income), Race (racialized vs white populations) and Race-Income (low-income racialized populations vs high-income white populations). Across all four measures of cannabis retail density, the most disadvantaged tracts (Q1) had a greater likelihood or odds of increased cannabis retail density, on average, compared to the most advantaged tracts (Q5). However, these differences varied based on density measure used. While the likelihood or odds of cannabis count per tract and cannabis retail density per population were not significantly different across quintiles of ICE comparing racialized vs white populations, quintiles with more racialized populations had increased odds of more cannabis retail density per square mile and per 10km of roadway. The most disparate effects were observed for cannabis retail density per square mile.

The purpose of Study 3 was to measure alcohol and tobacco retailer presence near cannabis retailers, using measures of proximity and density. The average distance from a cannabis retailer and the nearest alcohol or tobacco outlet was 495.5 feet (median: 130 feet). On average, there were 1.3 and 3.0 alcohol or tobacco outlets within 500 and 1000 feet of cannabis retailers on the census tract level. In the adjusted model, greater proportions of Black residents were associated with less distance between cannabis retailers and the nearest alcohol or tobacco outlet and suburban and rural tracts versus urban tracts were associated with more distance between outlets. Low-income tracts compared to high-income tracts and a greater proportion of vacant units were associated with a greater count of alcohol and tobacco outlets within 500 and 1000 circular feet of cannabis retailers. A greater percentage of Hispanic residents, and suburban and rural tracts vs urban tracts were associated with a lower count of alcohol and tobacco outlets near cannabis retailers.

In conclusion, the results of this dissertation suggest that, nationally, census tracts with the highest levels of neighborhood racial or socioeconomic disadvantage have greater cannabis retail density compared to more advantaged tracts. It was common for cannabis retailers to be located near alcohol and tobacco outlets, with some areas or populations subjected to structural disadvantage demonstrating a greater density or proximity of alcohol and tobacco outlets near cannabis retailers, suggesting that these disparate patterns of retail density may be related. In general, this dissertation is a novel line of inquiry with significant implications for research, surveillance, and evaluation efforts going forward. This research can inform how states and localities regulate cannabis retail location to ensure socially equitable outcomes in practice.